hamtaa-texttools 0.1.44__py3-none-any.whl → 0.1.46__py3-none-any.whl

{hamtaa_texttools-0.1.44.dist-info → hamtaa_texttools-0.1.46.dist-info}/METADATA

@@ -1,12 +1,12 @@
 Metadata-Version: 2.4
 Name: hamtaa-texttools
-Version: 0.1.44
+Version: 0.1.46
 Summary: A set of high-level NLP tools
 Author: Tohidi, Montazer, Givechi, Mousavinezhad
 Requires-Python: >=3.8
 Description-Content-Type: text/markdown
-Requires-Dist: openai>=1.97.0
-Requires-Dist: numpy>=1.26.4
+Requires-Dist: openai==1.97.1
+Requires-Dist: numpy==1.26.4
 
 # Text Tools
 

{hamtaa_texttools-0.1.44.dist-info → hamtaa_texttools-0.1.46.dist-info}/RECORD

@@ -1,4 +1,4 @@
-texttools/__init__.py,sha256=cI10Q_zaM9DPUCVOM79gZceuyt6Pjgpj3R-AG7xgUM8,778
+texttools/__init__.py,sha256=ZaopVfEJlp9n9mgtdYC0JcpRwAMuf99lUD8VsgemT20,784
 texttools/base/__init__.py,sha256=KUGm-Oe0BxlrRhPS-Jm2q1NCmwX8MdtZtloia7bcLaM,189
 texttools/base/base_categorizer.py,sha256=ojup94iXLxh92TjiJmrFXeRbsWKlon7PPAqez96B1bs,1130
 texttools/base/base_keyword_extractor.py,sha256=uKpxb3xI-sim-vXWe1R4_36QRhSNsWDR4IuVdpkZMME,868
@@ -11,9 +11,6 @@ texttools/base/base_router.py,sha256=pFDjIXFqAhPiS9Onu5py_GxOq8geDGJDQh6k6IhCkvw
 texttools/base/base_summarizer.py,sha256=7NAilhUPs6ZUwkBpTtXAj6n2XxQH1w6SOolf3gQX2gc,1627
 texttools/base/base_task_performer.py,sha256=3-6qshkie50S7pRG4WHRNC_RdUbSmHOPKW56CD92-rM,1852
 texttools/base/base_translator.py,sha256=BoOxqaoPoUs8t1O3m2yL9pQa5iwisl097immTVcGZoE,1020
-texttools/batch_manager/__init__.py,sha256=3ZkxA395lRD4gNxJ1vp0fNuz_XuBr50GoP51rrwQ0Ks,87
-texttools/batch_manager/batch_manager.py,sha256=jAmKskL3OTYwwsO1mWsWAB3VxMlOF07c2GW1Ev83ZhY,9283
-texttools/batch_manager/batch_runner.py,sha256=kW0IPauI11xpssApMA7b4XI19FePImywym3V7tBaa-o,7404
 texttools/formatter/__init__.py,sha256=KHz2tFZctbit_HVbQNCTMi46JzmKlg-uB6Ost63IpVU,46
 texttools/formatter/base.py,sha256=0fiM6E7NdJevAVpL6yyPaUZVJGKWxE3fr-Ay1oqgJqQ,879
 texttools/formatter/gemma3_formatter.py,sha256=AmdKBYLj6HMsI2DDX4KHNEEVYJmz_VVNUBOv8ScGjsY,1865
@@ -53,8 +50,12 @@ texttools/tools/summarizer/__init__.py,sha256=phrR7qO20CNhO3hjXQBzhTRVumdVdGSufm
 texttools/tools/summarizer/gemma_summarizer.py,sha256=ikhsBv7AiZD1dT_d12AyjXxojzSW92e2y5WjchI_3bE,4474
 texttools/tools/summarizer/llm_summerizer.py,sha256=-0rUKbSnl1aDeBfJ5DCSbIlwd2k-9qIaCKgoQJa0hWc,3412
 texttools/tools/translator/__init__.py,sha256=KO1m08J2BZwRqBGO9ICB4l4cnH1jfHLHL5HbgYFUWM8,72
-texttools/tools/translator/gemma_translator.py,sha256=57NMfJAZHQjZSr_eCBePE_Pnag8pu3O00Jicxhzn6Jc,7572
-hamtaa_texttools-0.1.44.dist-info/METADATA,sha256=OImC1zmuJh7p8SY3s3mhm8poOzYOuuqx6vjOeDy5O3k,1481
-hamtaa_texttools-0.1.44.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
-hamtaa_texttools-0.1.44.dist-info/top_level.txt,sha256=5Mh0jIxxZ5rOXHGJ6Mp-JPKviywwN0MYuH0xk5bEWqE,10
-hamtaa_texttools-0.1.44.dist-info/RECORD,,
+texttools/tools/translator/gemma_translator.py,sha256=rbP0kgkhOiEPdHWgHQc7Lev7lrAIYqNb6t_OfZLp44E,7180
+texttools/utils/flex_processor.py,sha256=Y44uTracvXUJiUm5hh57Uk0933RU9GTc3dN_1Bo_XQA,3214
+texttools/utils/batch_manager/__init__.py,sha256=3ZkxA395lRD4gNxJ1vp0fNuz_XuBr50GoP51rrwQ0Ks,87
+texttools/utils/batch_manager/batch_manager.py,sha256=jAmKskL3OTYwwsO1mWsWAB3VxMlOF07c2GW1Ev83ZhY,9283
+texttools/utils/batch_manager/batch_runner.py,sha256=kW0IPauI11xpssApMA7b4XI19FePImywym3V7tBaa-o,7404
+hamtaa_texttools-0.1.46.dist-info/METADATA,sha256=zZG-0IaOyeEFQO0rhBpT194Jsst-uuqRevcizIm8tiI,1481
+hamtaa_texttools-0.1.46.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
+hamtaa_texttools-0.1.46.dist-info/top_level.txt,sha256=5Mh0jIxxZ5rOXHGJ6Mp-JPKviywwN0MYuH0xk5bEWqE,10
+hamtaa_texttools-0.1.46.dist-info/RECORD,,

texttools/__init__.py

@@ -1,4 +1,4 @@
-from texttools.batch_manager import BatchJobRunner, SimpleBatchManager
+from texttools.utils.batch_manager import BatchJobRunner, SimpleBatchManager
 from texttools.handlers import (
     NoOpResultHandler,
     PrintResultHandler,

texttools/tools/translator/gemma_translator.py

@@ -1,5 +1,3 @@
-import json
-import re
 from typing import Any, Optional
 
 from openai import OpenAI
@@ -35,7 +33,7 @@ class GemmaTranslator(BaseTranslator):
         **client_kwargs: Any,
     ):
         super().__init__(handlers)
-        self.client = client
+        self.client: OpenAI = client
         self.model = model
         self.temperature = temperature
         self.client_kwargs = client_kwargs
@@ -134,27 +132,16 @@ class GemmaTranslator(BaseTranslator):
         messages.append({"role": "user", "content": text_prompt})
 
         restructured = self.chat_formatter.format(messages=messages)
-        completion = self.client.chat.completions.create(
+        completion = self.client.chat.completions.parse(
             model=self.model,
             messages=restructured,
-            response_format={
-                "type": "json_schema",
-                "json_schema": {
-                    "name": "NER",
-                    "schema": PreprocessorOutput.model_json_schema(),
-                },
-            },
+            response_format=PreprocessorOutput,
             temperature=self.temperature,
-            **self.client_kwargs,
+            extra_body=dict(guided_decoding_backend="auto") ** self.client_kwargs,
         )
-        response = completion.choices[0].message.content
-
-        # Remove Markdown-style triple backticks and any optional language tag like "json"
-        if response.startswith("```"):
-            response = re.sub(r"^```(?:json)?\s*|```$", "", response.strip())
-
-        entities = json.loads(response)
+        message = completion.choices[0].message
 
+        entities = message.parsed
         return entities
 
     def translate(
@@ -189,7 +176,7 @@ class GemmaTranslator(BaseTranslator):
             temperature=self.temperature,
             **self.client_kwargs,
         )
-        response = completion.choices[0].message.content.strip()
+        response = completion.choices[0].message.content
 
         self._dispatch(
             {

texttools/utils/flex_processor.py

@@ -0,0 +1,78 @@
+import random
+import asyncio
+from openai import OpenAI, RateLimitError, APIError
+from typing import Optional
+from pydantic import BaseModel, ValidationError
+import httpx
+
+# http_client = httpx()
+# test_client = OpenAI(http_client=http_client)
+
+async def flex_processing(
+        LLM_client: OpenAI,
+        system_prompt: str,
+        user_prompt: str,
+        output_model: Optional[BaseModel]=None,
+        prompt_cache_key: Optional[str]=None,
+        max_retries: int = 10,
+        base_delay: float = 2.0,
+        model_name: Optional[str] ="gpt-5-mini",
+        **client_kwargs):
+    """
+    Wrapper for flex processing with retry and exponential backoff.
+    Handles 429 'Resource Unavailable' errors gracefully.
+    """
+    for attempt in range(max_retries):
+        try:
+            request_kwargs = {
+                    "model": model_name,
+                    "messages": [
+                        {"role": "system", "content": system_prompt},
+                        {"role": "user", "content": user_prompt},
+                    ],
+                    "service_tier": "flex",
+                    "timeout": 900.0,
+                    **client_kwargs
+                }
+            if output_model:
+                request_kwargs["response_format"] = output_model
+            if prompt_cache_key:
+                request_kwargs["prompt_cache_key"] = prompt_cache_key
+            
+            response = LLM_client.chat.completions.parse(**request_kwargs)
+            # response = self.client.chat.completions.parse(output_model)
+            content = response.choices[0].message.content
+                # ✅ Validate structured output if a model is provided
+            if output_model is not None:
+                try:
+                    output_model.model_validate_json(content)
+                    base_content = response.choices[0].message.parsed
+                    # base_content = output_model(**content)
+                    return base_content
+                except ValidationError as ve:
+                    # Treat invalid output as retryable
+                    wait_time = base_delay * (2 ** attempt) + random.uniform(0, 1)
+                    print(
+                        f"[Flex Retry] Attempt {attempt+1}/{max_retries} produced invalid structured output. "
+                        f"Retrying in {wait_time:.2f}s... (ValidationError: {ve})"
+                    )
+                    await asyncio.sleep(wait_time)
+                    continue 
+        except (RateLimitError, APIError) as e:
+            wait_time = base_delay * (2 ** attempt) + random.uniform(0, 1)
+            print(
+                f"[Flex Retry] Attempt {attempt+1}/{max_retries} failed "
+                f"with error: {type(e).__name__} - {e}. "
+                f"Retrying in {wait_time:.2f}s..."
+            )
+            await asyncio.sleep(wait_time)
+
+        except Exception as e:
+            # Non-recoverable error: break out immediately
+            raise RuntimeError(
+                f"[Flex Processing] Unrecoverable error for prompt_key={prompt_cache_key}: {e}"
+            )
+
+    raise RuntimeError(
+        f"[Flex Processing] Exhausted {max_retries} retries for prompt_key={prompt_cache_key}"
+    )